Akshay K

I built FactShield AI, a hallucination detection system that verifies claims against trusted sources. It uses Groq API with Llama 3.3 70B, FAISS for semantic matching, and EasyOCR to extract text from images and documents.

I implemented a confidence-scoring pipeline to flag misleading AI outputs with explainable reasoning. The project won a competitive hackathon across AI and full-stack tracks.

I also built a real-time MERN stack chat application with JWT authentication and MongoDB-based message persistence. It includes RESTful APIs, Socket.IO messaging, and AI-powered message assistance using LLM APIs.

As an Artificial Intelligence & Machine Learning Intern at SmartBridge Educational Services Pvt. Ltd., I completed a virtual internship program and gained practical exposure to AI/ML concepts and workflows. At Students Tribe, I performed data annotation and labeling to support AI/ML model training and evaluation.